Universidade de Aveiro Departamento deElectr´onica, Telecomunica¸c˜oes e Inform´atica, 2016
S´
ergio Ant´
onio
da Costa Coelho
Mecanismos de tuning em redes de acesso ´
oticas
Universidade de Aveiro Departamento deElectr´onica, Telecomunica¸c˜oes e Inform´atica, 2016
S´
ergio Ant´
onio
da Costa Coelho
Mecanismos de tuning em redes de accesso ´
oticas
Tuning mechanisms in optical access networks
Disserta¸c˜ao apresentada `a Universidade de Aveiro para cumprimento dos requisitos necess´arios `a obten¸c˜ao do grau de Mestre em Engenharia Electr´onica e Telecomunica¸c˜oes, realizada sob a orienta¸c˜ao cient´ıfica do Prof. Dr. Ant´onio Teixeira e do Prof. Dr. M´ario Lima, ambos do De-partamento de Electr´onica, Telecomunica¸c˜oes e Inform´atica (DETI) e do Instituto de Telecomunica¸c˜oes (IT) da Universidade de Aveiro
o j´uri / the jury
presidente / president Prof. Doutor Nuno Borges de Carvalho
Professor catedr´atico da Universidade de Aveiro
vogais / examiners committee Professor Doutor Ant´onio Lu´ıs Jesus Teixeira
Professor associado da Universidade de Aveiro (orientador)
Prof. Doutor Pedro Tavares Pinho
agradecimentos Esta disserta¸c˜ao marca o final de um cap´ıtulo da minha vida. Foi um percurso moroso e dif´ıcil, com momentos altos e baixos, dos quais levo recorda¸c˜oes que ficaram comigo para o resto da vida. Aproveito para deixar alguns agradecimentos as pessoas que mais me marcaram neste percurso. Em primeiro lugar, n˜ao poderia deixar de ser, um agradecimento especial aos meus pais, Cristina e Nelo, aos meus irm˜aos, Patricia, Manuel e Margarida, pelo seu apoio incondicional. Certamente sem eles tudo isto n˜ao seria poss´ıvel. A minha afilhada, Luana, por ser uma fonte de alegria constante na minha vida.
Gostaria tamb´em de agradecer aos meus amigos de infˆancia, Marcel, C´atia, Virg´ınia, Carla, Joel e Coralie, fomos criados juntos e muito aquilo do que sou hoje devo tamb´em a vocˆes.
Gostaria de agradecer ainda a todos amigos e colegas de vida acad´emica, Carlos Salgueiro, Rui Puga, Samuel Moreira, Andr´e Faceira, Jo˜ao Gon¸calves, Tiago Urze, Hugo Veloso, Lu´ıs Pinheiro, Jo˜ao Andrade, Lu´ıs Matos e Miguel Leite tanto pelos momentos de divers˜ao como pelas noites de estudo. Todos v´os, de uma maneira ou de outra me ajudaram a moldar e expandir os meus horizontes, certamente me tornaram uma pessoa melhor. Um abra¸co especial ao Jos´e Maricato, companheiro de longa data tanto em Portugal como em Erasmus.
N˜ao podia deixar de mencionar tamb´em os, poucos mas bons, amigos que fiz em Erasmus: Mladen, Nikola, Bojan, David, Branka, Ida, Heinrich, Felipe, Milica.
Uma men¸c˜ao ainda para a Eng. Ana Tavares pela sua incans´avel ajuda no laborat´orio, um muito obrigado pela paciˆencia disponibilizada.
Por fim, um agradecimento especial aos meus orientadores, Prof. Dr. Ant´onio Teixeira e Prof. Dr. M´ario Lima, pela ajuda dispendida, mas principalmente pela motiva¸c˜ao que me deram para finalizar este trabalho
Palavras-Chave NG-PON2, TWDM-PON, PtP WDM-PON, redes de acesso, redes ´oticas passivas, lasers sintoniz´aveis, mecanismos de tuning, receptores sin-toniz´aveis, tempo de sintoniza¸c˜ao, precis˜ao de tuning.
Resumo O principal tema abordado neste trabalho ´e a tecnologia a ser utilizada em redes passivas ´oticas de nova gera¸c˜ao, nomeadamente o TWDM e o PtP WDM, com um foco em especial nos mecanismos de tunabilidade a ser usados nas arquiteturas de rede mencionadas. Come¸cou por ser feita uma abordagem geral ao tema, um ”overview” a recomenda¸c˜ao para o NG-PON2 ´e apresentada, assim como uma revis˜ao a transceivers incolores, componentes essenciais nas arquitecturas do NG-PON2. Tendo em conta o n´ıvel de precis˜ao apresentada, trˆes tipos de ONUs s˜ao definidas nos stan-dards do NG-PON2, e os mecanismos de tuning necess´arios para lidar com os diferentes tipos de ONU s˜ao tamb´em apresentados.
Foi caracterizado um laser sintoniz´avel (DFB) tendo em vista a sua uti-liza¸c˜ao numa ONU do NG-PON2, os parametros avaliados foram: tempo de sintoniza¸c˜ao, excurs˜ao espectral e precis˜ao de tuning. As t´ecnicas de medi¸c˜ao s˜ao apresentadas bem como os resultados obtidos.
Keywords NG-PON2, TWDM-PON, PtP WDM-PON, access networks, passive optical networks, tunable lasers, tuning mechanisms, tunable filters, tuning time, tuning accuracy.
Abstract The main issue addressed in this work are the technologies to be employed in the next-generation passive optical networks, including TWDM-PON and PtP WDM, with a particularly focus on the tuning mechanisms featuring the aforementioned network architectures. A general approach to the topics was carried out, by making an overview the NG-PON2 recommendation, a review to colorless transceivers is presented as well, essential components on the PON2 architectures. Three types of ONUs are defined in NG-PON2 standards, by taking into account the accuracy level of the ONU Tx, tuning mechanisms necessary to to deal with the different kind of ONUs are presented as well
A tunable DFB laser was characterized, considering its utilization on a NG-PON2 ONU. The evaluated parameters are: tuning time, spectral excursion and tuning accuracy. The setups utilized for the measurements are presented as well as the results.
Contents
Contents i
List of Figures v
List of Tables vii
List of Acronyms viii
1 Introduction 1
1.1 Context and Motivation . . . 1
1.2 Structure and Objectives . . . 2
1.3 Contributions . . . 3
2 State of Art 5 2.1 Telecommunication Network . . . 5
2.2 Optical Access Network . . . 6
2.3 PON and the FTTx Network Models . . . 7
2.4 Legacy PON Systems . . . 9
2.4.1 G-PON . . . 10
G-PON architecture . . . 10
G-PON trasmission and main features . . . 11
Security and Protection . . . 12
2.4.2 XG-PON/NG-PON1 . . . 13
2.5 Tunable Light sources . . . 13
2.5.1 Semiconductor lasers . . . 14
2.5.2 Distributed Feedback Laser Diode (DFB) . . . 14
2.5.3 Distributed Bragg Reflector laser (DBR) . . . 15
2.5.4 External Cavity Laser (ECL) . . . 16
2.5.5 Vertical Cavity Surface Emitting Lasers (VCSEL) . . . 16
2.5.6 Lasers based on Hybrid III-V on SOI (Silicon-on-insulator) . . . 17
2.6 Tunable Filters . . . 18
2.6.1 Microring resonator . . . 18
2.6.2 Silicon Fabry-Perot filter (FPF) . . . 18
2.6.3 Thin Film Filter (TFF) . . . 19
3 NG-PON2 General Requirements 21
3.1 System Overview . . . 21
3.2 Architecture . . . 23
3.3 Line Rates . . . 23
3.4 Wavelength Plan . . . 24
3.5 Advanced Network Functions . . . 25
3.5.1 Spectral Flexibility . . . 26
3.6 Coexistence with Legacy PONs . . . 26
3.7 ODN Considerations . . . 27
3.7.1 Wavelength Selected ODN . . . 28
3.7.2 Wavelength Routed ODN . . . 28
3.7.3 Hybrid ODN . . . 29
3.8 Key technologies . . . 30
3.8.1 Transmitters . . . 30
3.8.2 Receivers . . . 31
3.8.3 Minimum Tuning Window . . . 32
3.9 ONU Tunablity . . . 33
3.9.1 Wavelength Assignment and Tunning . . . 33
3.9.2 ONU Activation Process . . . 35
3.9.3 Tunning Time Classes . . . 36
Tuning Time . . . 36
3.9.4 Maximum Spectral Excursion (MSE) . . . 37
3.9.5 Wavelength Control . . . 38
3.9.6 Wavelength Stability . . . 38
3.9.7 Wavelength Accuracy . . . 38
3.9.8 Wavelength Locking . . . 39
3.9.9 OLT-port Protection . . . 40
3.9.10 OLT-port sleep and Load Balancing . . . 41
3.10 Externally Modulated Lasers versus Direct Modulated Lasers for ONU Tx transceiver . . . 42
3.11 Wavelength Multiplexers (WM) . . . 43
3.11.1 Cyclic Characteristics of AWG . . . 43
3.12 OLT Configuration Options . . . 44
3.13 Management and Control . . . 44
3.14 Conclusion . . . 45
4 Tunable Laser Testing 47 4.1 Introduction . . . 47 4.2 DFB Characteristics . . . 47 4.2.1 DFB Tuning . . . 49 4.3 Tuning Time . . . 50 4.3.1 Setup . . . 51 4.3.2 Methodology . . . 52 4.3.3 Results . . . 53 Scenario 1 . . . 53 Scenario 2 . . . 54 Scenario 3 . . . 55
4.3.4 Statistic analysis . . . 56
4.4 Tuning Accuracy . . . 58
4.5 Tuning Mechanisms . . . 59
4.6 Further Considerations . . . 61
5 Conclusions and Future Work 63 5.1 Conclusions . . . 63
5.2 Future Work . . . 64
Appendices 65
A Laser Characterization 66
B Tuning Time measurements 69
List of Figures
2.1 Architecture of a telecommunication network [2] . . . 5
2.2 High-level architecture of an access network [6] . . . 7
2.3 FTTx application model in PON [4] . . . 8
2.4 Passive optical network architecture [4] . . . 8
2.5 PON Evolution [8]. . . 9
2.6 Typical GPON architecture [9] . . . 10
2.7 G-PON wavelength allocation . . . 11
2.8 Downstream Transmission. . . 11
2.9 Upstream Transmission . . . 12
2.10 Allocation of G-PON XG-PON wavelengths . . . 13
2.11 G-PON and XG-PON coexistence . . . 13
2.12 DFB Structure array [19]. . . 15
2.13 DBR laser schematics [20] . . . 15
2.14 ECL Structure [26]. . . 16
2.15 VCSEL Structure. . . 17
3.1 NG-PON 2 system diagram [43] . . . 23
3.2 NG-PON 2 wavelength Plan [47] . . . 27
3.3 WS-OND with power splitter (top) and WR-ODN with lumped cyclic AWG (bottom). Relevant interface reference points are also shown [48] . . . 28
3.4 Hybrid ODN and WS-ODN in comparison [48] . . . 29
3.5 (a) 4 channel array EML (b) 4 channel array APD [51] . . . 32
3.6 Wavlength assignment/tunning message exchange between OLT and ONU [53] 34 3.7 Laser tunning across the operating band [47] . . . 37
3.8 Illustration of the Maximum Spectral Excursion [47] . . . 38
3.9 Result for a mistuned ONU, a) Negative correction required, b) Positive cor-rection required. [56] . . . 39
3.10 N:1 Protection Architecture [57]. . . 40
3.11 Exemple of OLT-port sleep. (a) Normal state (b) Normal state with less traffic (c) sleep state [49]. . . 41
3.12 a) Optical spectrum of a 10 Gb/s EML Tx measured with a resolution band-width of 2.5 GHz. b)Gaussian-shaped filter and optical spectrum of a 10 Gb/s DML Tx measured with a resolution bandwidth of 2 GHz and tuned to fit within MSE [56] . . . 42
3.13 Cyclic AWG US wavelength routing, the colors represent the four different CT. [56] . . . 43
3.14 Example of OLT configuration [56] . . . 44
4.1 DFB laser . . . 48
4.2 Setup used to characterize the DFB laser . . . 48
4.3 Wavelength and Optical Power vs I bias @ 25◦C . . . 49
4.4 Wavelength and Optical Power vs I bias @ 40◦C . . . 49
4.5 Wavelength vs I bias vs @ room temperature, 35◦C, 40◦C and 50◦C . . . 50
4.7 Tunning Time measurement setup (top), Expected Results (bottom) . . . 51
4.8 Different scenarios for a change from λ1 yellow to λ2 green . . . 52
4.9 Oscilloscope picture of λ1 yellow, and λ2 green, 20 ms/div . . . 53
4.13 (a)theat and (b) tcool ms/GHz plotted with mean value and mean value ± standard deviation. . . 56
4.14 theat/GHz histogram and normal distribution . . . 56
4.15 tcool/GHz histogram and normal distribution . . . 57
4.16 Tuning time measurement, 20 ms/div . . . 58
4.17 Optical Spectrum of the DFB, externally modulated at 2.5 Gbit/s, and mea-sured with 33 pm accuracy. . . 59
4.18 Relationship between I heater and frequency shift @ 25◦C. . . 59
List of Tables
2.1 Characteristics of the several types of a telecommunication network [6] . . . . 6
2.2 GPON Nominal bit rate [9]. . . 11
2.3 DBR Tunable Lasers [21] [22] [23] [24] [25]. . . 16
2.4 Categories of tunability . . . 18
3.1 NG-PON2 Line Rate [5]. . . 24
3.2 NG-PON2 Wavelength Plan [5]. . . 24
3.3 Network function and Tuning Frequency/Time [46] . . . 26
3.4 Classes for Optical Path Loss in NG-PON2 [5] . . . 28
3.5 Possible Tunable transmitters in the ONU [49] . . . 30
3.6 Possible solutions for ONU Rx [49] . . . 31
3.8 Minimum Tuning window for TWDM-PON ONU transmitters [5] . . . 32
3.7 Use cases for with Tx and Rx solutions presented in tables 3.5 and 3.6 . . . . 33
3.9 Wavelength channel tuning time classes for NG-PON2 [47] . . . 36
3.10 Maximum Spectral Excursion allowed in NG-PON2 . . . 37
3.11 N:1 Protection Steps [57]. . . 40
4.1 Tuning time summary Scenario 1 . . . 54
4.2 Tuning time summary Scenario 2 . . . 55
4.3 Tuning time summary Scenario 3 . . . 55
4.4 theat/GHz and tcool/GHz measurements with TEC . . . 57
A.1 λ versus I bias, I heater=0 @ 25◦C . . . 66
A.2 λ versus I bias, I heater=0 @ 45◦C . . . 66
A.3 λ versus I heater, I bias=50.66 mA @ 25◦C . . . 67
A.4 λ versus I heater, I bias=80.2 mA @ 25◦C . . . 67
A.5 λ versus I heater, I bias=50.13 mA @ 40◦C . . . 67
A.6 λ versus I heater, I bias=80.14 mA @ 40◦C . . . 68
List of Acronyms
AES Advanced Encryption
Standard
AMCC Auxiliary Management and Control Channel
AN Access Network
APON Asynchronous Transfer Mode-based Passive Optical Network
ATM Asynchronous Transfer Mode
AWG Arrayed Waveguide Grating BPON Broadband Passive Optical
Network
CAPEX Capital Expenditure
CE Coexistence Element
CG Channel Group
CO Central Office
CS Channel Spacing
CT Channel Termination
DBA Dynamic Bandwidth
Allocation
DCT Dispersion Compensation Technique
DBR Distributed Bragg Reflector
DFB Distributed Feedback
DML Direct Modulated Laser DWBA Dynamic Wavelength and
Bandwidth Allocation EML Externally Modulated Laser
FEC Forward Error Correction FSAN Full Service Access Network FSR Free Spectral Range
FTTB Fiber to the Building FTTC Fiber to the Curb
FTTH Fiber to the Home
IP Internet Protocol
ITU International
Telecommunication Union NG-PON2 Next-Generation Passive
Optical Network 2 OLT Optical Line Terminal
ONU Optical Network Unit
OPEX Operating expense
OSA Optical Spectrum Analyzer
P2P Point to Point
P2MP Point to Multi-point
PAYG pay-as-you-grow
PLOAM Physical Layer Operations, Administration and
PON Passive Optical Network PtP WDM Point-to-Point
Wavelength-division multiplexing
QoS Quality of Service
RN Remote Node
Rx Receiver
SOA Semiconductor Optical Amplifier
SSW Single-Side Spectral Width
SW Wavelength Channel
Selector
TEC Thermo-Electric Cooler
TFF Thin Film Filters
TDMA Time Division Multiple Access
TDM-PON Time-Division Multiplexing Passive Optical Network
Tx Transmitter
TWDM-PON Time and Wavelength Division Multiplexed Passive Optical Network
US Upstream
WM Wavelength Multiplexer
WR-ODN Wavelength Routed ODN WS-ODN Wavelength Selected ODN XG-PON1 10×Gigabit Passive Optical
Chapter 1
Introduction
Nowadays we live in an era that might be called connected age. In contrast with its predecessor, information age in which the core business was and still is to develop monolithic intangible goods and profiting using licensing fees and strict control of software copying, connected age comes in the picture when enterprises started monetizing the human behavior on the web. [1] With the evolution and rise of technologies such as smart phones and tablets, and with the proliferation of network access points, to be or not connected is in the tip of your finger. Thus social networking became a huge thing, from Facebook to Youtube, and even Twitch TV (streaming service), all services that profit with the connectivity of people. Also the availability to buy and download products from the web (music, films, games) cause a ever increasing demand of network bandwidth, therefore the network it self must evolve in order satisfy the users needs, new future services and maintain a certain Quality of Service (QoS).
1.1
Context and Motivation
When it comes to select an ideal solution for an Access Network (AN) there is a continuous struggle to balance infrastructure investments against the need to deliver stronger financial returns. It is with this factors in mind that Passive Optical Networks (PONs) for access come in the picture.
In a PON, a single fiber between the Optical Line Terminal (OLT) and the Remote Node (RN) is shared by all users connected to it. The network between the OLT and the Optical Network Units (ONUs) is passive, which means that there is no need of any power supply in this path.
PON standardization started in the early 1990s, by the hands of the Full Service Access Network (FSAN) working group, and led to the creation of APON in 1995, which later was improved to Broadband Passive Optical Network (BPON) in 1998, redefining it in 2005 to allow higher bit rates. The maximum data rates achievable for this technology are 1.2 Gb/s and 622 Mb/s for downstream and upstream respectively. [2] [3]
In 2001, FSAN group started the development of GPON (Gigabit Passive Optical Net-work), standardized in 2003 by the International Telecommunication Union (ITU), to keep pace with the increasing demand for bandwidth. This later standard supports downstream bit rates up to about 2.5 Gb/s and upstream bit rates up to about 1.25 Gb/s. [2] [4]
All currently deployed PON based on the standards referred before are known as Time-Division Multiplexing Passive Optical Networks (TDM-PONs) architectures since they rely
on TDM technology. Upstream transmission is accomplished by time sharing the available bandwidth between all subscribers (TDMA), while downstream operation is performed by sending all data to all ONUs, being these responsible to select the data destined to the subscriber(s) associated. [2] [4]
Due to progressive massification of new types of traffic (such as 3DTV and cloud-based services) the ITU has already ratified new standards. 10×Gigabit Passive Optical Net-work (XG-PON1) was standardized in 2010. It allows 10 Gb/s and 2.5 Gb/s bit rates for downstream and upstream transmissions, respectively. Next-Generation Passive Optical Net-work 2 (NG-PON2) specifications are being standardized, the first set of recommendations was issued in the start of 2013, G.989.1. [5], allowing 40 Gb/s symmetrical bandwidth, using a Time and Wavelength Division Multiplexed Passive Optical Network (TWDM-PON) based architecture, with a Point-to-Point Wavelength-division multiplexing (PtP WDM) PON over-lay.
TWDM-PON increases the aggregate PON rate by stacking four XG-PON1 using four pairs of wavelengths. The only significantly new components in TWDM, and used for sim-ple network deployment and management purposes, are colourless tunable transmitters and receivers.
The NG-PON2 specifications for wavelength plan, and channel spacing are covered by the well matured tunable devices technologies developed for core/metro networks, (full C-band or L-band with more than 80 channels). However, they require a considerably reduced number of wavelength channels. Although there is a need for low volume of TWDM downstream transceivers and costs will be shared across multiple subscribers, when it comes to choose an upstream Transmitter (Tx) low cost solutions are preferable. Taking this into consideration, NG-PON2 specifications defines three types of ONU Tx [5]: calibrated, loosely calibrated and uncalibrated. Therefore multiple tuning mechanisms must be enabled in the PON in order to activate and deal with the different kinds of ONUs.
NG-PON2 systems takes advantage of the tunable Tx and Rx, not only they provide colorless ONUs and pay-as-you-grow deployment, some advanced network functions resulting from in-service wavelength tuning techniques may feature in the system as well. The maximum service interruption required for in-service tuning in NG-PON2 is 50 ms [5] Therefore tuning time is an important ONU parameter, which will be decisive regarding the implementation or not of in-service tuning functionalities.
1.2
Structure and Objectives
This document is divided in 5 chapters, and the main objectives are the following: • Overview of the NG-PON2 recommendation.
• Study the available tunning mechanisms in NG-PON2 systems. • Overview on tunable transceivers.
• Testing of tunable lasers available in the laboratory, regarding the tunable characteris-tics.
In the first chapter are presented the context and motivation, structure and objectives, as well as the contributions according to the author’s opinion.
The second chapter presents the state of the art, in that sense PON’s preponderance in access networks is reviewed as well as the Legacy PONs structures and main features in order to introduce the NG-PON2 systems.
In the third chapter an overview of the NG-PON2 recommendation is presented, with a focus in the tunable characteristic of the network architecture. The relevant parameters of tunable ONUs are briefly described and a quick overview on tunable lasers’ tuning mechanisms and receivers is also presented.
Fourth chapter features a characterization of tunable transceivers to be used in NG-PON2 as well as the experimental results on a transceiver tuning time measurement and accuracy.
Finally in the fifth chapter conclusions taken from the carried out work are presented as well and some future work related to this topic is suggested.
1.3
Contributions
In the author’s opinion the main contributions of this work are as follows:
• Identification of the currently available tuning mechanism to be used in NG-PON2 deployment, distinguishing three types of tuning mechanisms regarding the use of cali-brated, loosely calibrated and uncalibrated transmitters.
• Characterization of a tuning transmitter to be used in NG-PON2, tunable time mea-surement, and determination of tuning accuracy for a DFB laser.
Chapter 2
State of Art
2.1
Telecommunication Network
A modern telecommunication network general structure consists of three main sub net-works: backbone or core network, metro or regional network, and the access network. Fig. 2.1 presents a simple scheme of a telecommunication network. The Backbone Network as the core, interconnects all metro/regional networks of a country. The Metro Networks aggre-gate the high traffic from access networks’ (AN) Central Office (CO), transport the data at higher speed, pass that traffic addressed to other metro/regional areas to the core network and delivers to the respective central office the remaining traffic.
The core and metro networks structures are usually more uniform that the access network, and have their cost shared by a large number of users. Finally, the access network which provides end-user connectivity, e.g. AN is the bridge between the user and the service provider.
Table 2.1: Characteristics of the several types of a telecommunication network [6] Network Type Characteristics
Core Has active elements;
Low granularity;
Low variation in traffic flow;
Operate with a lower number of protocols than metro networks;
Transport traffic over long distances; Use a irregular mesh topology; Metropolitan Has active elements;
Medium granularity;
Medium fluctuation in traffic flow; Covers high density population areas; Use ring topologies;
Access Passive optical network (PON); High granularity;
High fluctuation in traffic flow;
Operate with a variety of protocols (IP, ATM, Ethernet);
Covers small distances;
Have a wide variety of topologies implemented
2.2
Optical Access Network
Nowadays the most cost efficient solution in access networks is the PON, which makes the fiber access one of the most important technologies in the next generation networks. Before overlooking the role of PON’s in AN, is of the most importance to understand what an AN is. An AN provides the connection between subscribers and service providers, it is the last segment of an operator network which enables users to access the network itself, either to transmit or receive data, voice and so on. Optical AN are constituted of an optical line terminator (OLT) located at the CO, Remote Node (RN), and a Optical Network Unit (ONU) to terminate the fiber.
Figure 2.2: High-level architecture of an access network [6]
An AN can be perceived as two distinct networks, the distribution and the feeder network, as can be seen in Fig.2.2 The first one connects the RN with the respective ONUs, and the second one connects the CO with the RNs. The RN can be active or passive.
2.3
PON and the FTTx Network Models
According to Ovum [7] PON is, nowadays, the dominant FTTx technology due to the cost advantages of its point-to-multipoint architecture compared to other solutions. Ovum also forecasts that more than 90% of the expected 300 million FTTH subscribers in 2019 will be PON Networks. Fig. 2.3, represents the several applications models of an optical access networks, the so called FTTx. The different architectures are characterized by the fiber length and the ONU location relatively to the user end [4]. Citing, Fiber to the Curb (FTTC), Fiber to the Building (FTTB) and Fiber to the Home (FTTH).
Figure 2.3: FTTx application model in PON [4]
PONs are the attractive solution for the FTTx due to the inherent characteristics of its passive RN, Fig. 2.4 represents the architecture of an FTTH PON, in contrast, and active RN requires constant power supply, backup power, and additionally, a cabinet for its placement, raising the Capital Expenditure (CAPEX) and Operating expense (OPEX).
PON is a Point to Multi-point (P2MP) system, as it is referred before. Some authors argue that P2P optical network can be considered a PON due to connection between the OLT and the ONU being completely passive. Nevertheless the most common and widespread definition to PON is P2MP.
The feature of passiveness makes the network deployment flexible and advantageous [6]: • Both, equipment and fiber in the CO are shared (smaller number of fibers in feeder),
providing lower cost than in a Point to Point (P2P) solution. Above all, power supply is not needed for the RN, greatly reducing the CAPEX;
• Trough the removal of the active equipment, there is a reduction in the electromagnetic interference, decreasing the failure rate of the line and the external equipment. This provides easy maintenance, lower CAPEX and also lower OPEX;
2.4
Legacy PON Systems
ITU Study Group 15 Q2 completed the recommendations that define a GPON (gigabit passive optical network) [ITU series G.948] in 2004, since then, FTTx based on PON have been largely deployed. Full Service Access Networks (FSAN), along with the ITU are the forum and the standardization entities with the greatest activity in study this type of networks. Their point of view on next generation networks divides the evolution on two stages, NG-PON1 and NG-PON2, the first one is known as 10×Gigabit Passive Optical Network (XG-PON1) and it is considered as short-term solution, a transition technology, standardized since 2010, while NG-PON2 technology is seen as a medium-term solution (2015) Fig. 2.5, its standardization was initiated on the beginning of 2013 [5].
2.4.1 G-PON
Gigabit passive optical network (G-PON) is currently a mature technology, ITU started working on G-PON in 2002 and there is the G.984.x series of recommendations, standard series define general characteristics of GPON (G.984.1) as well as physical layer specification (G.984.2), transmission layer specification (G.984.3) and ONU (Optical Network Unit) man-agement and control specification (G.984.4). G-PON has an enhanced capability comparing with its predecessors, APON and BPON (ATM PON which evolved in to Broadband PON), and is backward compatible. Being G-PON a passive network, only the OLT and the ONU are active equipments.
G-PON architecture
Sarting from the Central Office (CO), only one single-mode fiber strand is connected to a splitter near the users’ end, Fig.2.6.
Figure 2.6: Typical GPON architecture [9]
The optical splitter simply divides the signal into N separate paths to the subscribers. From this point, the number of paths that the optical power can be divided into may vary from 2 to 64, thus the split ratio capability of a G-PON network is 1:128. After the splitting, individual single-mode fiber strand run to each user (home, businesses, etc.). The optical fiber transmission span from the central office to the each user can be up to 20 km [9].
The network is capable of supporting the bandwidth requirements of business and residen-tial services and covers systems with lot of different line transmission rates for downstream and upstream direction.
Table 2.2: GPON Nominal bit rate [9]. Transmission Direction Bit Rate
Downstream 1244.16 Mbit/s 2488.32 Mbit/s Upstream 155.52 Mbit/s 622.08 Mbit/s 1244.16 Mbit/s 2488.32 Mbit/s G-PON trasmission and main features
The operating wavelength range is 1480-1500 nm for the downstream direction and 1260-1360 nm for upstream direction. In addition, the wavelength range 1550-1560 nm can be used for downstream RF video distribution, Fig.2.7.
Figure 2.7: G-PON wavelength allocation
Downstream transmission, on G-PON, consists in broadcasting fixed frames of 125µs from the OLT to the ONUs. Although all the ONUs receive the same data, GEM port ID are imbued in the frames which allows them to be differentiated therefore a filtering process will occur in the ONUs, a respective ONU will only receive data that was aimed to it in the first place [10] [11].
Figure 2.8: Downstream Transmission. [12]
(TDMA), Time Division Multiple Access in order to avoid packet collision. Each ONU has a respective time window when they are allowed to transmit upstream data. An Upstream Bandwidth Map is sent in the downstream frames which regulates time attribution to each ONU [10] [11].
Figure 2.9: Upstream Transmission [12]
Features such as Forward Error Correction (FEC) and Dynamic Bandwidth Allocation (DBA) are enabled .FEC is a mathematical signal-processing technique that encodes data so that errors can be detected and corrected. DBA is a methodology that allows quick adoption of users bandwidth allocation based on current traffic requirements [10] [13] [14].
Security and Protection
Regarding security, in the upstream direction G-PON uses point-to-point connection so all traffic is secured from eavesdropping. When it comes to downstream, the basic function-ality of G-PON is that data is broadcast to all ONUs and every ONU have allocated time when data belongs to it (TDM). This could present a problem, since a malicious user could reprogram his own ONU to receive all downstream data belonging the other ONUs connected to that respective OLT. To prevent the referred security issues the GPON recommendation G.984.3 describes the use of an information security mechanism to ensure that users are al-lowed to access only the data intended for them. Advanced Encryption Standard (AES) is encryption algorithm that is used, it accepts 128, 192, 256 byte keys which makes the en-cryption extremely hard to compromise. Thus this key can be changed periodically, without disturbing the information flow, to enhance security [10].
There are two types of protection switching, Automatic switching and Forced switching. The first one is triggered by fault detection, such as loss of signal, loss of frame, signal degrade and so on. The second one is activated by administrative events, such as fiber re-routing, fiber replacement, etc. However, protection is considered as an optional mechanism because its implementation depends on the realization of economical systems.
2.4.2 XG-PON/NG-PON1
XG-PON (10-Gigabit-capable passive optical network) is a technology already standard-ized in the ITU Recommendations G.987.x, since 2010, and defines a mechanism migration to acquire transmissions of 10Gbit/s downstream and 2.5Gbit/s upstream. As it was referred before, XG-PON is viewed as a short-term solution, therefore, there are few technological suppliers, few interoperable equipment between manufacturers and few interested operators. XG-PON can be seen as the natural evolution of the G-PON Networks, nevertheless the need for larger bandwidth will lead operators to evolve directly into NG-PON2 [8], since its stan-dardization is expected within this year, there are already the ITU Recommendation G.989.1 which defines the general requirements as well as physical layer specification G.989.2.
Wavelength plan is defined in ITU Recommendation G.987, Fig.2.10
Figure 2.10: Allocation of G-PON XG-PON wavelengths [8]
The operating wavelength range is 1575-1580nm for the downstream direction and 1290-1330nm for upstream direction. Coexistence of XG-PON with the G-PON technology is possible within the same fiber that requires a WDM filter in the CO to combine the users signal and the RF-Video [10] [13] [15], Fig.2.11.
Figure 2.11: G-PON and XG-PON coexistence [8]
2.5
Tunable Light sources
Semiconductor lasers have the potential to meet demands for next generation high speed optical networks. Their low cost, wide tunability, low power consumption, and very good spectral response make them ideal transmission sources. Colorless components will be es-sential in NG-PON2 deployment. In the this section we will review some characteristics of
semiconductor lasers, in particular the tuning mechanisms behind such devices that provide a colorless mode of operation within a determined wavelength range.
2.5.1 Semiconductor lasers
Semiconductor lasers are popular optical communication light sources for high speed data transmission. They are accepted as “the laser of the future”, due to their compactness, easy integration, and high output powers. Coherent emission is produced in these lasers by stimulated emission, and the gain is achieved in the active medium of semiconductor by electrical injection [16]. Their gain material is usually a compound of elements from columns III and V of the periodic table. Typically they operate in the 1310-nm or 1550-nm region of the spectrum.
These diode lasers are compact, therefore, they are developed on a large scale using very mature fabrication technologies. Semiconductor lasers are very efficient in converting electrical power into optical power [17]. In order to study such type of device is important to first understand the basic structures and operations that comprise semiconductor lasers. From the simplest diode laser to the most complex one, all structures function on the same principle, an active medium (gain medium) is placed between two mirrors (Distributed Bragg Reflectors, High and Anti-reflector coatings or distributed grating), while the active medium is stimulated by injecting a current in order to produce emission, the mirrors provide a feedback and a selective function. When stimulated emission occurs, light from unwanted modes is reflected into the cavity in order to stimulate more photons. In short their wavelength is determined both by the properties of the III-V gain medium and by the physical structure of the laser cavity surrounding the gain medium [18].
2.5.2 Distributed Feedback Laser Diode (DFB)
In the DFB laser diode the active medium is periodically structured as a diffraction grating, figure 2.12 (bottom). The grating povides feedback and its wavelength (corrugation period) determines the wavelength emitted from the laser. The tunablity in this type of laser is achieved by thermally change the reflection index of the grating, thus this type of device is tunable in a narrowband, typically under 5nm. A conventional DFB laser’s emission wavelength can be tuned by heating or cooling with a gradient of about 0.08 nm/K. Widely tunable devices using DFB technology are available by incorporating various laser cavitys in the same device, DFB arrays, figure 2.12, coarse tuning is achieved by selecting the correct laser bar, and fine tuning is done thermally.
Figure 2.12: DFB Structure array [19].
2.5.3 Distributed Bragg Reflector laser (DBR)
The DRB laser structure lies within the placement of at least one Bragg Reflector outside the gain medium (active region) at one end of the resonator, and the other end may have another DBR, Figure 2.13a shows the schematics of a simply DBR laser. The gratings provides a wavelength-dependent feedback and thus a single-mode operation can be achieved. Although DBR lasers are usually a single-frequency laser, they present a small tunable range within the operating wavelength, this tuning is associated with the thermal characteristics of the Gratings, the reflection index of the grating can be changed by driving it with a current. Moreover a tuning within the free spectral range of the laser resonator may be accomplished with a separate phase section, which can be electrically heated, or simply by varying the temperature of the gain region via the drive current, wider tunning can be achieved by adding more sections, figure 2.13b.
(a) 2 section DBR Laser (b) 3 section DBR Laser
Figure 2.13: DBR laser schematics [20]
DBR Tunable Lasers are electrically tunable, e.g., different modes are locked by injection current into one, or more existing sections of the laser, furthermore the lasing wavelength is a
function of a multi-reflection-peak resultant of the feedback provided by the one or more DBR mirrors inside the cavity, lasers using this technology, such as, SG-DBR [21], SSG-DBR [22], GCSR [23], DS-DBR [24], DCG-DBR [25], have been widely reported in literature. In Table 2.3, some parameters and characteristics regarding this type of laser are presented.
Table 2.3: DBR Tunable Lasers [21] [22] [23] [24] [25].
Laser-Parameter Output Power Tuning range SMSR Structure Tuning Currents Digital Supermode DBR
(DS-DBR) ∼14 dBm 45 nm >40 dB
Front and Rear Gratings, Gain and Phase Section.
Front (0 - 200 mA) Rear (0 - 60 mA) Sampled Grating DBR
(SG-DBR) ∼4 dBm 38 nm >40 dB
Front and Rear SGDBR mirrors,
Lever, Gain and Phase Sections. N/A Super Structure Grating DBR
(SSG-DBR) ∼9 dBm 40-60 nm >35 dB
Rear and Front Reflectors, Gain and Phase Sections.
Front (0 - 15 mA) Rear (0 - 18 mA) Digital Concatenated Grating DBR
(DCG-DBR) ∼14 dBm 60 nm >30 dB
Rear and Front Concatenated Gratings,Gain and Phase Sections.
Front (0 -30 mA) Rear (0 -40 mA) Grating assisted codirectional
Coupler with rear Sampled grating Reflector (GCSR)
∼14 dBm 40 nm >30 dB
Rear Reflector, and Coupler in the Front Section, Gain and Phase Section
Front (0 - 16 mA ) Rear (0 - 16 mA)
2.5.4 External Cavity Laser (ECL)
The ECL is based on a laser diode where the cavity is completed with external optical components and typically as one Anti-reflection coating end. The external components can vary, as well as the setup their set, nevertheless the external setup is made of a set of lens in order to provide a collimated beam, and some kind of diffraction grating in order to provide wavelength selectivity. Wavelength tuning is possible by including and adjustable element, a mirror or a diffraction grating, usually by mounting it on a structure such as a piezoelectric crystal. Figure 2.14 shows two types of configuration, a) Littrow configuration, b) is Littman–Metcalf configuration, the first one exhibits better output power, since the beam is reflected only once, in the other hand the second offers better wavelength selectivity.
Figure 2.14: ECL Structure [26].
This type of structure provides a very wide tuning range, and a very narrow linewidth, but a relatively slow tunning.
2.5.5 Vertical Cavity Surface Emitting Lasers (VCSEL)
VCSEL achieve single longitudinal mode operation in different manner, the structures presented before are edge emitting lasers while the VCSEL a surface emitting laser. The cavity is placed between two highly reflective Bragg Reflectors (99.5% on the emitting surface
and 99.9% on the other surface), one of the advantages of VCSEL lies within its structure, which provides a easier way to couple the emitted light into fiber. In the VCSEL tunability is achieved by including Micro-electro-mechanical (MEMs) part in the structure, the lasing wavelength is therefore selected by a combination of MEMs and thermal Tuning, figure 2.15.
Figure 2.15: VCSEL Structure.
VCSEL/MEMs structures present relatively wide tunning, 30nm, but there is one concern in using MEMs, that lies within their long term mechanic reliability.
2.5.6 Lasers based on Hybrid III-V on SOI (Silicon-on-insulator)
Hybrid III-V on SOI devices integrates the advantage of III-V efficient light emission and amplification with the low loss and high index contrast waveguiding, for efficient spectrum filtering and tuning, of the silicon. The studied devices, are presented in [27, 28], use a set of extra-cavity ring resonators (RRs), and Bragg Reflectors (optional) in order to form two resonating structures with slightly different free spectral range (FSR), the RRs can thermally tuned, and the difference in the RR’s FSR allows taking advantage of the Vernier Effect for wavelength selectivity and tunning. Also the FSR difference of the rings will determine the tuning range of the device. Such devices present wide tuning range 40 to 45 nm, and high values of SMSR, above 40 dB, and narrow linewidth, in the order of a few hundred kilohertz. Similar devices may be realized using at set of extra-cavity ring resonators on a Fabry-Perot cavity.
Based on the presented content above, it can be concluded that semiconductor lasing is associated with laser’s cavity length, and intrinsic properties of the semiconductors used in the active sections. These properties rule the single-wavelength lasing characteristics of a laser. As far as tunability goes, three distinct categories can be distinguish: Temperature, Current Driven and Mechanical. The three are not mutually exclusive, since there are devices
where a combination of at least two types are used. Table 2.4 summarizes the characteristics of the different tunning categories.
Table 2.4: Categories of tunability
Tuning Mechanism Wavelength Range Tuning Speed
Temperature NarrowBand 5nm Slow (sub-s order)
Mechanical (MEMs, piezoelectric transducer) WideBand ECL - 100nm VCSEL- 30nm
Very Slow (s-order) Current Driven Wideband - >30 nm Fast (ns to µs order)
2.6
Tunable Filters
In NG-PON2, a tunable wavelength selection device is required in the ONU, because each ONU should be able to get access to all wavelengths for dynamic bandwidth allocation. The wavelength selective filter is required to have wide tuning range to maximize the number of channels that can be selected, negligible crosstalk to avoid interference from adjacent channels, fast tuning speed to minimize the access time, small insertion loss, polarization insensitivity, stability against environmental changes (humidity, temperature, vibrations, etc.), and last but not the least, low cost [18]. Microring-resonator, thin film filter, Bragg gratings, Bragg reflectors, liquid crystal tunable filter, thermally tuned and angle tuned FP filters are the research focus. Several techniques will be introduced as example.
2.6.1 Microring resonator
Thermally tunable Si microring resonator (MRR)filter with flat-top passband is a promis-ing candidate for WDM signal processpromis-ing, because it is compact and easy to be integrated with other Si devices, including electro-optic modulators, optical routers and reconfigurable optical add drop multiplexers (ROADMs), etc. A thermally tunable Si 3rd order MRR filter has been proposed, which shows a box-like response with low intra-band ripple ( 0.65 dB), low insertion loss (less than 0.9 dB) and out-of-band rejection higher than 40 dB. By applying power to TiN heaters, the filter can be thermally tuned over the whole feedback shift register (FSR) of 3 nm with the tuning efficiency of 48.4 mV/nm [29].
2.6.2 Silicon Fabry-Perot filter (FPF)
FPF, also called as etalon, is realized by a silicon wafer with both sizes carefully polished. The incoming light experiences a multiple-beam interference. As a result, the wavelengths that match the resonant conditions are selected at the filter output as indicated in Eq 2.1
λm= (2nLccos θ)/m, (2.1)
where n is the etalon reflective index, Lc is its length. By varying the reflective index or
the length of the FP resonator, we can tune the output wavelength of the FP filter. For the reflective index tuning, we can use thermo optic effect, which is the refractive index variation induced by a temperature change. The thermo-optic effect is very strong and the thermally induced expansion of a silicon wafer is very poor, resulting in a large pre valence of the
thermo-optic effect. A thermal tuning coefficient of 0.083nm/K is obtained at 1550 nm. The device can also be tuned by changing the optical path length in the cavity, which may be achieved mechanically, thermally, or electro-optically. Hence, the FP filter is a versatile and flexible component and a prime candidate for a wavelength selective filter in TWDM-PON systems. Furthermore the main advantage of fiber FP filters is that they can be integrated within the system without incurring coupling losses.
2.6.3 Thin Film Filter (TFF)
As a group, micro electro mechanical systems (MEMS) Fabry-Perot devices tend to possess wide tuning range, but have an serious limitation. They are structurally restricted to the simplest type of single-cavity etalon design. Plasma enhanced chemical vapor deposition (PECVD) is the preferred process for dense, compliant, homogeneous optical coatings of thin film silicon [30]. It is known that thin film narrowband filters may be tuned by mechanical rotation of the angle of incidence [31]. Nonetheless TFF including an integrated heater for thermal tuning were presented in [32]
2.6.4 Fiber Bragg grating (FBG)
FBG is also a technology to keep in mind due to its sharp spectral filtering character-istics and low insertion loss. Conventionally, the Bragg wavelength of FBGs can be shifted by imparting either the strain or temperature effect on a uniform FBG. Temperature change linearly shifts the Bragg wavelength through the refractive index variation of the grating induced by the thermo-optic effect [33]. Mechanically induced strain physically extends or compresses the grating period, which leads to a linear shift of the Bragg wavelength in con-junction with the refractive index change in the grating through the strain-optic effect. The excellent tolerance of silica-glass fibers to mechanical strain makes strain tuning preferable for achieving a wide tuning range.
Tunable receivers and tunable transmitters have been a research topic of optical transport networks for more than a decade, therefore there is a great deal of development practice in this area. The NG-PON2 system applications takes advantage of this component development effort in a couple of ways. Frist, the NG-PON2 tunable transceivers reuse mature tunable optical transport network components, if one technology does not perform to expectations, there is always other options to provide the required function. This versatility reduces the risk component availability. Second NG-PON2 can provide significant relief on the specifications of tunable optical transport network components. Because the NG-PON2 wavelength tuning performance could be relaxed from that of the optical transport network and NG-PON2 channel rates are widely used in the optical transport network, critical tuning requirements, such as wavelength tuning range, tuning speed, channel spacing, can be dramatically relieved. Such performance relaxation offers significant yield improvements during the mass production and cost reductions for tunable transceivers. [34]
Chapter 3
NG-PON2 General Requirements
The next generation passive optical network stage 2(NG-PON2) project was initiated by FSAN community in 2011. [35] The project’s goal is to investigate optical fiber networks technologies enabling a bandwidth increase beyond the XG-PON’s 10Gbit/s in the access network.Next-Generation Passive Optical Network 2 (NG-PON2) is a multifaceted system, capable of meeting the different needs of a wide range of networks in a diverse market, and is deployable in several applications in a efficient manner. Taking in consideration that the ODN represents 70% of sum of the investments in the PON roll-out, NG-PON2 will not be a disruptive solution since re-utilizes the ODN infrastructure from previous PONs. Characteristics from legacy PON systems and ([ITU G.982], [ITU G.983], [ITU G.984], [ITU G.987]) are maintained as much as possible, as well as the re-use of established technical capabilities, to ensure some backward compatibility with the already exiting ODN’s.
Many PON technologies have been proposed to provide broadband optical access beyond 10 Gb/s. There are the 40 Gigabit time-division multiplexed PON (40G TDM-PON) pro-posal [36] which increases the single carrier serial downstream bit rate of a 10 Gigabit PON (XG-PON) [15] to 40 Gb/s, the time- and wavelength-division multiplexed PON (TWDM-PON) proposal which stacks multiple XG-PONs using WDM [37], a group of WDM-PON pro-posals which provide a dedicated wavelength channel at the rate of 1 Gb/s to each ONU with different WDM transmit or receive technologies [38] [39] a set of orthogonal frequency-division multiplexed (OFDM)-based PON proposals which employ quadrature amplitude modulation and fast Fourier transform to generate digital OFDM signals for transmission [40] [41]. How-ever, the requirement for backwards compatibility blocked of the way for WDM-PONs because they require wavelength selective ODNs. 40G TDM-PON is also out of consideration due to the cost pressure of 40G components in each user and the fiber chromatic dispersion would greatly limit transmission distance [42]. Other options are eliminated due to either techni-cal immaturity or time-frame requirement. As result, all attentions turned to TWDM-PON, which attracted the majority support from global vendors and was selected among all of the aforementioned proposals by the FSAN community in April 2012 as a primary solution for NG-PON2, with PtP WDM overlay channels.
3.1
System Overview
• Multiple wavelength channel TWDM architecture.
• 4-8 TWDM channel pairs (each channel pair comprising one downstream and one up-stream wavelength channel), configurable for incremental growth starting from one de-ployed channel pair (i.e., not all channel pairs need to be active); for example, “pay as you grow” capability of TWDM populating in the OLT.
• Downstream and upstream nominal line rates per channel: – 10 Gbit/s downstream and 10 Gbit/s upstream – 10 Gbit/s downstream and 2.5 Gbit/s upstream – 2.5 Gbit/s downstream and 2.5 Gbit/s upstream
• Passive fiber reach of at least 40 km and maximum differential fiber distance up to 40 km with configurable maximum differential fiber distance as 20 km and optionally 40 km,
• Capability to reach 60 km, preferably with passive outside plant, • Support for a split ratio of at least 1:256.
NG-PON2 systems require flexibility to balance trade-off in speed, distance, and split ratios for numerous applications. The set of parameter combinations that are supported by the system must include:
• 40 Gbit/s downstream capacity and 20 km reach with at least 1:64 split ratio, • 10 Gbit/s upstream capcity and 20 km reach with at least 1:64 split ratio, • Access to peak rates of 10/2.5 Gbit/s downstream/upstream,
• Longer distances with lower split ratios are also possible. NG-PON2 systems may also support:
• 40 Gbit/s downstream capacity with 10 Gbit/s per upstream channel and 20 km reach with at least 1:64 split ratio,
• 2.5 Gbit/s per downstream channel and 2.5 Gbit/s per upstream channel with 40 km reach with at least 1:32 split ratio,
• 10 Gbit/s per downstream channel and 10 Gbit/s per upstream channel with 40 km reach with at least 1:32 split ratio,
• Acces to peak rates of 10/10 Gbit/s downstream/upstream,
3.2
Architecture
The Basic NG-PON 2 architecture is shown in Figure 3.1. TWDM-PON increases the ag-gregate PON rate by stacking four XG-PON1 using four pairs of wavelengths (e.g, wavelength pairs of {λ1↑,λ1↓},{λ2↑,λ2↓},{λ3↑,λ3↓} and {λ4↑,λ4↓} in Fig.3.1). A TWDM-PON system with four pairs of wavelengths is able to provide 40Gbit/s downstream and 10Gbit/s upstream with 40 km reach and 1:64 split ratio. Each ONU can provide peak rates of 10Gbit/s and 2,5Gbit/s downstream and upstream, respectively, meeting the NG-PON2 requirements.
Figure 3.1: NG-PON 2 system diagram [43]
The PtP WDM overlay allows NG-PON2 to meet demanding operator requirements for business and back haul services. On the PtP WDM each ONU is served by a dedicated wavelength channel. In the baseline configuration, eight channels of PtP WDM are considered to allow full co-existence with legacy systems. Furthermore, depending on the particular deployment scenario, unused spectrum may be dedicated to additional PtP WDM channels.
3.3
Line Rates
Table 3.1: NG-PON2 Line Rate [5].
TWDM Downstream Line Rate (Gbit/s) Upstream Line Rate (Gbit/s)
Basic Rate 9.95328 2.48832
Rate Option 1 9.95328 9.95328
Rate Option 2 2.48832 2.48832
PtP WDM Downstream/Upstream Line Rate (Gbit/s)
Class 1 1.2288–1.2500
Class 2 2.4576–2.6660
Class 3 9.8304–11.09
Class 4 6.144
3.4
Wavelength Plan
As might be expected, with each generation of PON, the availability of unallocated spec-trum diminishes. Co-existence with legacy PONs and RF Video relies on the NG-PON2 wavelength plan. Some considerations were made regarding the elaboration of two distinct wavelength plans, considering the presence or not of RF Video. However the idea did not move forward, a single wavelength plan is a more attractive approach since it would mean a world wide standard.
The Upstream bands fo TWDM (table 3.2) are located in the C-band where there is high availability in components which will facilitate lower costs in the ONUs. TWDM down-stream is allocated in the L-band, although technology is not so matured, there is a need for low volume of TWDM downstream transceivers and costs will be shared across multiple subscribers.
Table 3.2: NG-PON2 Wavelength Plan [5].
TWDM Upstream Wavelength Band (nm) Downstream Wavelength Band (nm)
Wide 1524-1544
Reduced 1528-1540 1596-1603
Narrow 1532-1540
PtP WDM Downstream/Upstream Wavelength Band (nm)
Shared 1603-1625
Expanded 1524-1625
It may be noted that there are three upstream wavelength band options available, that were specified for TWDM-PON. These options are motivated by differing capabilities of the ONU transmitter to control its wavelength. A variety of techniques may be employed in order to control the wavelength in the upstream direction from the ONU to OLT. Neverthe-less, temperature is the primary parameter of control. The wide option may be used by a ”Wavelength-set” [44] approach to channel control, where the transmitter is allowed to drift in wavelength over a wide range, as the OLT de-multiplexer has cyclic pass bands. As for the narrow option a more precise transmitter is needed, since it will be of the utmost importance for it to precisely lock onto an assigned DWDM wavelength.
(CS) of 100Ghz from 187.1Thz to 187.8Thz. In the upstream direction CS of 50, 100 and 200GHz are to be supported, although no channel plan has been recommended it can be assumed that the ONU transmitter will adapt to whatever grid is employed by the OLT de-multiplexer.
It can be noticed as well that two bands were specified for the PtP WDM channels, the reasoning behind this fact will be further explained. The Shared Spectrum option, as it’s name implies, is the wavelength plan to be applied in a scenario of full coexistence with legacy PON systems (G-PON, XG-PON1 and RF Video). The Expanded Spectrum option exploits the concept of spectral flexibility of NG-PON2 by enabling unused bands, in any particular deployment, to be utilized by PtP WDM. The Expanded Spectrum option may also be the option to go to in a Greenfield scenario, with no legacy coexistence limitations.
3.5
Advanced Network Functions
The wavelength agility of NG-PON2 systems allows network functionalities that were previously unavailable in Legacy TDM-PONs. By exploiting the wavelength agility inherent to the tunability of the NG-PON2 ONUs new use case scenarios are opened up. The attractive network functions resulting from the use of wavelength tunability in services are:
• Incremental upgrade of the system capacity, or pay-as-you-grow (PAYG)
• OLT power saving
• Load Balancing among OLT ports
• OLT-port protection
• Flexible bandwidth assignment such as Dynamic Wavelength and Bandwidth Allocation (DWBA)
Furthermore, a NG-PON2 architecture featuring the functionalities listed above as already been demonstrated [45]. Table 3.3 summarizes the relationship between network functions , tuning frequency and the required tuning time. Some of this function will be addressed in further sections.
Table 3.3: Network function and Tuning Frequency/Time [46]
Function Description Tuning Frequency Tuning Time Colorless ONU
Eliminates complicated inventory management and improper connections between a colored ONU and an OLT.
Low (only initialization) Long (s order) Advanced Power Saving (OLT-port sleep)
For example, when there is less traffic, all ONUs are connected to one OLT-port and the other ports can be forced to sleep.
Medium (per hour or day)
Medium (ms order)
OLT-port Protection
When an OLT-port as failed, ONU can continue its communication to other OLT-ports by changing its wavelength.
Low (OLT-port Failure) Short (ns - ms order) Pay-as-you-grow (Incremental upgrade)
After installation of ONUs, the total bandwidth can be increased
incrementally by istalling additional OLT-ports.
Low (only initialization)
Long (s order)
Load Balancing
When the traffic of one OLT-port is heavily congested, some ONUs are assigned to communicate to other vacant ports.
Low (per day or week)
Medium (ms order)
Flexible Wavelength Assignment
The optimized wavelength and timeslot are assigned to each ONU in order to improve the usage efficiency of the wavelength and provide bandwidth according to user demand (DWBA).
High Short (ns - µs order)
3.5.1 Spectral Flexibility
Spectral Flexibility is one of the key features introduced in NG-PON2, it will facilitate a diverse range of deployment scenarios, network applications and evolution paths. In a practi-cal sense, and as it was explained before, whenever a particular subset of optipracti-cal spectrum is unused by TWDM and/or Legacy PON systems PtP WDM may make use of that particular sub-band.
Such flexibility can facilitate the support of different customer types on the same ODN in a flexibly way. Additionally, spectral Flexibility facilitate a range os system coexistence scenarios and allows operators to ”re-use” new wavelength bands when legacy systems are decommissioned.
3.6
Coexistence with Legacy PONs
The wavelength plan of NG-PON2 will facilitate its coexistence with Legacy PON systems. Regarding XG-PON1, NG-PON2 predecessor, the system uses wavelength overlay in order to permit each system to operate independently in a common fiber infrastructure. Not only simultaneous coexistence with legacy PON systems is achieved, but as well with 1555nm RF Video, figure 3.2 shows the NG-PON2 wavelength plan with inclusion of G-PON, XG-PON1 and RF Video.
The TWDM-PON downstream channels fit between XG-PON1 downstream and the mon-itoring band (labeled OTDR). On the other hand the TWDM-PON upstream channes work in the C-band which is expected to deliver lower cost optical transceivers at the ONUs.
The Shared Spectrum wavelength band for PtP WDM is shown in Figure 3.2, and it is configured as a mixed downstream and upstream plan, which will be assigned according to the network operator requirements.
Figure 3.2: NG-PON 2 wavelength Plan [47]
3.7
ODN Considerations
One of the biggest challenges in NG-PON2 deployment relates to operator requirements on the ability to re-use already installed PON fiber-infrastructure, therefore NG-PON PMD has been designed to be compatible with power splitter based ODNs. In order to re-use power-splitter based ODNs, already deployed for Legacy PON systems, the optical path losses supported by NG-PON2 transceivers must also be compatible.
Furthermore, the need for coexistence with Legacy PONs driven the NG-PON2 standards to assign the same nomenclature and values of the optical path loss used in XG-PON1. These values and classification of ODN can be seen in Table 3.4, and as it can be seen, the requirement for compatibility with a 15dB ODN loss has been maintained. On further notice, the 40Km fiber distance from the OLT to the furthest and closest ONU respectively is also supported on NG-PON2.
Table 3.4: Classes for Optical Path Loss in NG-PON2 [5]
ODN Class N1 N2 E1 E2
Minimum Loss (dB) 14 16 18 20 Maximum Loss (dB) 29 31 33 35
On a scenario where NG-PON2 needs to comply with Legacy PONs, and thus provide backward compatibility, the support of power-split only is mandatory. Two classes of PtP WDM ODN architectures are described in G.989: one requires tunable filters as the wave-length selection device, on the ONU, thus called Wavewave-length Selected ODN (WS-ODN). The other one, Wavelength Routed ODN (WR-ODN) uses wavelength splitters in the ODN, which allow a wavelength routing capability, figure 3.3. Nevertheless, the two are not mutually ex-clusive as PtP WDM ODN can be made as a combination of both, providing a hybrid solution which can take advantage of both configurations.
Figure 3.3: WS-OND with power splitter (top) and WR-ODN with lumped cyclic AWG (bottom). Relevant interface reference points are also shown [48]
3.7.1 Wavelength Selected ODN
NG-PON2 systems support single-stage splitting as well as multi-stage splitting, the re-quired support for the maximum split ratio is of at least 1:64. For coexistence with legacy PON systems, a wavelength-band multiplexer, which is referred to as Coexistence Element (CE), enables them to be combined in a single ODN. Maximum Reach can be traded off against split ratio, since the power loss using splitter devices scales with the total split ratio.
3.7.2 Wavelength Routed ODN
NG-PON2 systems entirely based on a WR-ODN only make sense in a Green-field de-ploymente scenario, when the PtP WDM makes use of the Expanded Spectrum configuration. Due to the lower insertion-loss of the wavelength splitters compared with power splitters, WR-ODN can provide either a longer reach or the usage of transceivers with lower budget classes.
Note that NG-PON2 systems need a Wavelength Multiplexer (WM) in order to combine multiple wavelength channels into the same ODN. The WM is considered an integrated com-ponent of the OLT equipment, in that sense the extra loss from the WM is not part of the ODN loss, which is specified from the output of the WM (the S/R-CG reference point) to the input of the ONU (R/S reference point). S/R-CG refers to the location where the OLT sends/receives a set of DS/US wavelength channel pairs, called Channel Group (CG), to/from ONUs. At the OLT, a logical Channel Termination (CT) function terminates each individual TWDM or PtP WDM channel.
The WM, and the wavelength splitters for WR-ODN may be realized with Arrayed Waveg-uide Grating (AWG) or Thin Film Filters (TFF)
3.7.3 Hybrid ODN
As it was mentioned before, an NG-PON2 system can comprise a Hybrid configuration of WS and WR ODN types. Such approach, in comparison to WS-ODN, can provide a higher split ration for the same OPL Classes. Example shown in Figure 3.3, is relative to a two stage power splitting ODN with a 26-28 dB loss budget. Replacing the second stage power splitter with an AWG, results in a 16-18 dB loss budget, allowing a more reasonable power budget [48]. Other possible configuration would be to use a wavelength splitter in the first stage, and grouping wavelengths as wavelength sets on the second stage, all wavelengths are sent to the ONUs as a single set via power splitter. [44]
Figure 3.4: Hybrid ODN and WS-ODN in comparison [48]
In short, NG-PON2 can support a choice of ODNs, i.e, WS-ODN, WR-ODN or even a combination of both. WR-ODN offers the advantage of reduced splitting loss that may be used in order to achieve extended system reach or allowance for the use of lower specification transceivers, and is more attractive towards rural areas and Green-field scenarios. Also, some services could make use of the physical layer channel isolation offered by wavelength splitting to offer virtual private line services using wavelength channels. Whereas WS-ODN has a stronger affinity toward (dense) urban ares, in particular if legacy ODNs are already in place.
3.8
Key technologies
3.8.1 Transmitters
The NG-PON2 wavelength plan defines the C-band and L-band for US and DS respec-tively. Therefore the expected specifications for tunable Transmitter (Tx) in the ONUs are 4-channel tunability, respecting the chosen CS (50 GHz, 100 GHz or 200 GHz) in the C-band, a line rate of 2,5 Gbit/s (most likely for residential service) or 10 Gbit/s (most likely for business and mobile services), and burst-mode operation.
These specifications are covered by the well matured tunable devices technologies devel-oped for core/metro networks, some of them were previously addressed in section 2.5 (full C-band or L-band with more than 80 channels). However, they require a considerably reduced number of wavelength channels and burst-mode capability [49]. Table 3.5 shows examples of possible tunable Tx to be employed in a NG-PON2 ONU.
Table 3.5: Possible Tunable transmitters in the ONU [49]
Type A A’ B C Device Heater-Integrated DFB-LD EML (TEC inside) DBR-LD (short cavity type) 4h EML array + Selector (SW) Tuning
Control Thermal Thermal Electrical Electrical (SW)
Tuning
Time sub-s order sub-s order
<ns (LD
chip level) <100 ns Modulation Direct mod. External
mod. Direct mod.
External mod. DCT for
10 Gbit/s 3 7 3 7
Type A is a conventional uncooled direct modulated DFB, but needs an integrated heater and temperature sensor for an on chip tunabilty [50].
Advantages: • Small size. • Low cost.
• Easy burst-mode operation.
Disadvantages: • Long tuning time.
• Needs Dispersion Compensation Tech-nique (DCT) for 10 Gbit/s operation. Type A’ is similar to the DFB previously mentioned, has an integrated modulator and Thermo-Electric Cooler (TEC) inside the module.
Advantages: • Small size. • Low cost.
• Does not need DCT, due to external modulation.
Disadvantages: • Long tuning time
Type B is a Distributed Bragg Reflector (DBR), one problem with this laser is that might suffer from mod-hop problem, but one solution is to adopt a short cavity design [?]
Advantages:
• Very short tuning time. • Small size.
Disadvantages
• Needs DCT for 10 Gbit/s operation. • More costly than previously solutions. Type C is a composed of a 4-channel arrayed device, an optical multiplexer, and an electric Wavelength Channel Selector (SW), figure3.5 [51]
Advantages:
• Short tuning time. • Mode hop free.
• Does not need DCT, due to external modulation.
Disadvantages: • Bulky Size.
• Complex configuration for monolithic integration.
3.8.2 Receivers
Similarly to the T-Tx, tunable receivers in the ONU are expected to have a 4-channel tunability, with 100GHz channel spacing (CS for the NG-PON2 DS) in the L-band and line rates of 2.5 and 10 Gbit/s. Implementing Tunable Rx in NG-PON2 represents one of the major challenges in the system deployment. It will be the first time that a tunable Rx will be implemented in a practical manner in the telecom field [49], which is an indicator that tunable Rx technology is less mature than tunable Tx. Table 3.6 shows examples of possible technology solutions for ONU Rx.
Table 3.6: Possible solutions for ONU Rx [49]
Type X Y Z Device Heater-integrated TFF + PD DeMUX + 4ch APD array + selector SW DeMUX + SOA gates + PD Tuning
Control Thermal Electric SW Optical Selector SW Tuning
Time s order <20 ns <1.5 ns (chip level)
Type X consists of a heater-integrated TFF pin-PD. The TFF is tuned by applying a current to the integrated heater, which shifts the central wavelength of the TFF. Thin film filters come in a small package, and offer a low cost solution, however they present a long tuning time, sub-s order.
Type Y is similar to Type C Tx, is composed of a 4-channel APD-arrayed device, an optical demultiplexer (DeMUX) and a SW, as shown in figure [51]. This Rx, was confirmed to have short tuning time <20 ns [49].
Type Z consists of a DeMUX, Semiconductor Optical Amplifier (SOA) gates, and a PD [52]. The SOA gates act as a wavelength selector, selecting the designated wavelength among the demultiplexed wavelengths with high speed tuning time of 1.5 ns. Although the use of the DeMUX is the same as the previous device, Type Z offers wavelength selection in the optical domain, while Type Y selects the wavelength at the electrical domain.
Figure 3.5: (a) 4 channel array EML (b) 4 channel array APD [51]
Table 3.7 presents which type of solutions can be achieved by employing the aforemen-tioned Tx and Rx in an NG-PON2 system. 7 marks the Tx and Rx that would be incompatible regarding the tuning time, from operation point of view does not make sense to have a fast Tx and a Slow Rx or vice versa.
3.8.3 Minimum Tuning Window
The minimum tuning window refers to the difference between the highest and the lowest operating frequencies/wavelengths of a tunable device, achievable by means of tuning control. Table 3.8 shows the minimum tuning window required for TWDM-PON ONU transmitters in NG-PON2 systems considering 50Ghz, 100Ghz and 200Ghz channels spacing and four and eight wavelength channel pairs. Regarding the PtP WDM-PON architecture in NG-PON2 systems, the minimum tuning window is based on the operating wavelength band.
Table 3.8: Minimum Tuning window for TWDM-PON ONU transmitters [5] Minimum Tuning Window Channel Spacing 4 channels 8 channels
50 Ghz 250 Ghz 450 Ghz
When using Cyclic Channel Grid 100 Ghz 500 Ghz 900 Ghz
200 Ghz 1000 Ghz 1800 Ghz
50 Ghz 175 Ghz 375 Ghz
When not using Cyclic Channel Grid 100 Ghz 340 Ghz 740 Ghz
